In this paper, we demonstrate how new technology from Oracle can be utilized to provide big data analytics acceleration in a streamlined fashion. Specifically, our approach leverages the acceleration capabilities of the Data Analytics Accelerator (DAX) unit provided by Oracle's T7/M7/S7 SPARC processors and the Java Stream API to seamlessly accelerate Java applications by up to 20X, while using drastically fewer resources.

In this paper we investigate the performance of cryptographic operations, when used in Java applications. We demonstrate the advantage of using built-in hardware accelerator for cryptographic operations on SPARC servers. In particular, we demonstrate the advantage of hardware cryptographic instructions invoked via AES and SHA intrinsics, implemented in the Java Virtual Machine (JVM), over the more traditional Java Native Interface (JNI) calls. For the purpose of our study, we modified the SPECweb2005 benchmark by adding modern banking requirements, and created a new workload which we call the End-to-End Java Security (EEJS) workload. Using the workload, we compare different Java Cryptographic Service Providers (CSPs) and arrive at the conclusion that hardware cryptography has significant performance advantage for Java applications. With the EEJS workload, we also identify several enhancements applicable to the Java Secure Socket Extension (JSSE).

This paper presents an operating system managed die-stacked DRAM called i-MIRROR that mirrors high locality pages from off-chip DRAM. Optimizing the problems of reducing cache tag area, reducing transfer bandwidth and improving hit latency altogether while using die-stacked DRAM as hardware cache is extremely challenging. In this paper, we show that performance and energy efficiency can be obtained by software management of die-stacked DRAM, which eliminates the need for tags, the source of aforementioned problems. In the proposed scheme, the operating system loads pages from disks to die-stacked DRAM on a page fault at the same time as they are loaded to off-chip DRAM. Our scheme maintains the pages in off-chip and die-stacked DRAM in a synchronized/mirrored state by exploiting the parallel loading capability to die-stacked and off-chip DRAM from the disk. This eliminates the need for physical page movement to the slower off-chip DRAM upon eviction from die-stacked DRAM. Requests for pages that got evicted from die-stacked DRAM are simply serviced by the slower off-chip DRAM to prevent frequent data movements of large pages and thrashing between conflicting pages. The operating system periodically monitors the usage of the pages in off-chip DRAM and promotes high locality pages to die-stacked DRAM. Our evaluations show that the proposed hardware-assisted software-managed i-MIRROR scheme achieves an IPC improvement of 13% while consuming 6% less energy than prior state-of-the-art die-stacked caching schemes and 79% improvement in terms of IPC and 72% in terms of energy savings over systems without die-stacked DRAM support.

We describe two SIP benchmarks developed by SPEC. SPECsip_Infrastructure2011 is a SIP Proxy benchmark with realistic user behavior modeling. The second benchmark, not yet officially named, is for a JSR-289 conforming application server. We describe the design, usage and the underlying methodologies for both benchmarks.

Many business-oriented services will be gradually offered in the Cloud. Java Message Service (JMS) is a critical messaging technology in Java-based business applications, particularly to those that are based on the Java Enterprise Edition (Java EE) open standard. Maintaining high performance in the horizontally scaled, and elastic, cloud environment is critical to the success of the business applications. In this paper, we present practical considerations in optimizing JMS performance for the cloud deployment, where some of the findings may also serve to improve the design of JMS container so it adapts well to cloud computing. Our work also includes performance evaluation on the proposed strategies.

There has been an increasing need of highly predictable, timely, and dependable communication services with QoS guarantees on an end-to-end basis either for embedded real-time applications or for multimedia-integrated distributed control. Performance objectives used in conventional networks-such as maximizing the throughput, minimizing the response time, or providing fairness to users-are not of the most important concern for both types of applications. Instead, resources should be appropriately reserved and managed to support multidimensional quality of service (QoS) on an end-to-end basis, as well as application-specific tradeoffs among them. The main intent of the paper is thus to address and demonstrate an environment-an integrated set of network resource management techniques, middleware layers, and network software-for supporting multi-dimensional QoS on an end-to-end basis in distributed real-time environments. The authors first provide an analytic QoS framework Then, they elaborate on the algorithms and mechanisms to provide such network services. To empirically analyze the behavior of the proposed components, they are currently implementing them as software layers on top of the Sun Solaris operating system, and comment on the implementation status.

A routing scheme is presented for a class of mesh topologies, which include tori and variations of tori. The variations model torus networks that have undergone topological perturbations. Such perturbations are due to addition and deletion of nodes, which may occur because of network failure and incremental network growth. Since the routing scheme is operable over a class of topologies, where addition or deletion of a node keeps a topology within the class, it is robust to incremental changes in network topology. In addition, it requires only minimal space and processing at the network nodes, which makes it suitable for high-speed implementation. Its efficiency, in terms of path length taken by packets, is demonstrated through both computer simulation and mathematical analysis. The analysis uses a random topology model to show that for “typical” topologies the scheme routes packets along near-optimum paths. © 1996 John Wiley & Sons, Inc.

The irregular torus, a class of topologies proposed for metropolitan area networks, is torus-like, and thus more reliable than the bus and ring topologies. There is a routing algorithm for the topologies which is space and time efficient so that it is suitable for high-speed implementations. In addition, it is robust to topology perturbations due to network growth and node/link failure. A buffer management is proposed to prevent store-and-forward deadlock for routing in the irregular torus, and the throughput of a typical irregular torus network is analyzed. Store-and-forward deadlock is avoided by a buffer management whose storage requirement and control complexity at a node are both constant with the size of a network. The throughput of a typical irregular torus network is shown to be within a constant factor of that of a square torus. Hence for a large number of nodes, it is much higher than the throughput of a bus, ring or tree.>